Report to Congress on the Costs and Benefits of Federal Regulations
Appendix: Summary of Public Comments
Quality of Cost and Benefit Estimates
The first issue on which we asked for comments in the July 22nd draft report concerned "The validity and reliability of the quantitative and qualitative measures of the costs and benefits of regulations in the aggregate, as well as of the individual regulations issued between April 1, 1996, and March 31, 1997, discussed in the attached draft report."
One theme that was repeated in several comments was that we had under emphasized the value of "knowing the exact amounts of total costs and benefits." These commenters argued that the public has a right to know what the total costs and benefits of regulations are and that regulatory accounting would help inform the political process and shape the debate over regulatory reform. They also note that our current position is inconsistent with statements made by OMB in the early 1990's that discussed the regulatory budget concept and by academic experts such as Hahn and Litan (1997) and Crandall, et al (1997).
It appears that there may have been some misunderstanding of our present and past position on the significance of aggregate data. It is true that we state ". . . but knowing the exact amounts of total costs and benefits, even if that were possible, adds little of value." But the operative terms are "exact" and "total." The fact that we do not believe that knowing the exact total cost or benefit of all regulations adds much policy content, does not mean that we believe that knowing the costs and benefits of individual regulations is unimportant. On the contrary, a key theme of the report is the importance for good regulatory decision making of knowing the costs and benefits of individual regulations before they are issued.
Nor is our position inconsistent with our past statements or with the academic articles cited on the subject of a regulatory budget. The past OMB statements cited were discussion pieces that recommended further study and experimentation with the concept of a regulatory budget. We still believe that further analytical work needs to be done to improve regulatory accounting and recommend such work in this report. But as Hahn and Litan (1997) state in their report, Improving Regulatory Accountability, which several commenters cite, ". . . aggregate estimates are less important initially because the best way to affect policy in the short term is to modify individual regulations . . ." Moreover, although they also add that ". . . aggregate estimates do have value," they are clearly referring to aggregate estimates for regulatory "areas," not the aggregate of all regulations. Even the regulatory budget they suggest Congress may wish to experiment with (See Crandall et al, 1997) is limited to specific areas and circumstances, such as where costs exceed benefits.
A second concern expressed by commenters about our estimates of costs and benefits was that our estimates were systematically biased. While the charge was leveled by a number of commenters, there was disagreement among them as to how our estimates were biased. Most industry representatives and think tanks asserted that our estimates understated costs and overstated benefits, while the public interest groups and Federal agencies generally asserted that our estimates overstated costs and understated benefits. Two reasons that were given for the belief that our estimates of costs and benefits were systematically biased warrant some examination.
First, is the theory that agency estimates, upon which many but not all of our estimates are based, systematically understate costs and overstate benefits because agency self-interest lies in regulation. Although this view of agency behavior enjoys widespread support among academics as a theoretical matter, there is little documentation available to support it -- perhaps because there are several potentially offsetting factors. For example, much of the data that agencies use to make their estimates of costs comes from the regulated entities who generally have the opposite incentives -- namely, they will likely overstate costs to help convince decision makers not to issue the regulation. Also, as noted in our report, competitive firms over time frequently find more cost-effective ways, including new technologies, to comply with regulations than had been envisioned ex ante. Some commenters pointed to a set of case studies that is about to be published to support this contention. On the other hand, there is a large body of literature that shows that agencies tend to overestimate the benefits of their programs because, over time, technological progress -- in communications to energy exploration to infectious disease -- has reduced the long run expected benefits of earlier regulations. We believe that the question of these types of systematic bias is unresolved and would profit from further analysis, which we intend to pursue over the next year or so.
The second reason that some commenters thought our cost and benefit estimates were systematically biased was because they claimed we ignored indirect costs and benefits. As we stated in the report, our totals included only direct cost and benefit estimates because there is no consensus about or transparent estimate of the indirect costs, and even if there were, there are no comparable estimates of the indirect benefits of regulation, which could occur because of the improved health and longevity that health, safety, and environmental regulation produce. We believe that the question of indirect costs and benefits also deserves further consideration.
Numerous commenters also raised questions about some of our specific estimates for economic, environmental and other social regulation. With respect to economic regulation, several commenters raised questions about our estimates of the costs and benefits of this type of regulation. They were especially concerned with our statement in the draft report:
"Economic regulation, including antitrust, may produce social benefits when natural monopolies are regulated to stimulate competition or when firms are prevented from anticompetitive collusion and mergers. In a dynamic economy, however, the dollar amount of such economic efficiency benefits are thought to be small (Hahn and Hird 1991)."Based on the Hahn and Hird survey, we entered "zero" for the benefits of economic regulation in the draft report. The comments that we received on this point convinced us that was a mistake. In this report we enter an asterisk for benefits with an explanation of what we had intended.
Part of the confusion that arose on this point stems from the fact that we were using a narrower definition of "economic regulation" than what may generally be understood to be included by the term. We stated in the draft report that:
"Economic regulation is so-called because it directly restricts firms' primary economic activities, e.g., its pricing and output decisions. It may also limit the entry or exit of firms into or out of certain specific types of businesses."Some commenters, on the other hand, thought economic regulation included anti-trust enforcement, financial soundness and capital market regulation, and regulation aimed at fraud and deception, which they assert provide significant benefits. We are partly to blame for this confusion because we did not make it clear that these types of activities, which may be viewed by some to be regulating economic activity, were not intended to be included in the "economic regulation" category because they do not directly regulate firms' pricing, output, or entry decisions.
For example, anti-trust enforcement by the Department of Justice and the Federal Trade Commission is not generally done through regulation. Financial soundness regulations, such as the Federal Reserve System's reserve requirements and deposit insurance together with the supervision and regulation required to avoid moral hazard, as well as SEC regulation of the capital markets, were also not counted in our estimates of either costs or benefits because they are not generally considered traditional economic regulation. If they were, the benefits of such "regulation" would almost certainly dominate the totals. As several commenters noted, these types of requirements may have averted a financial collapse such as we had during the Great Depression before these controls were put in place. Finally, the treatment of the regulation of fraud and deception is somewhat more complicated. To the extent it is done through general disclosure requirements, such as those issued by the FTC, the SEC and the Federal Reserve (truth in lending), it probably should be included in the social regulation category rather than in economic regulation because disclosure requirements do not regulate prices, output, or entry. In any case, the studies we used did not, for the most part, include either cost or benefit estimates of these types of activities. In our final report, we attempt to make it clear how we define " economic regulation," and we note that even for the activities we consider traditional regulation there are significant uncertainties. We did add a category for consumer protection regulation and for that purpose we used estimates of paperwork burden, based on our information collection budget for the independent agencies whose regulations we do not review. That provided us with cost estimates, but it does not provide benefit estimates. We again used an asterisk to indicate that significant unestimated benefits may exist from these disclosure requirements.
Several commenters also raised questions about the lack of benefits for economic regulations that were included in the Hahn and Hird estimates. The case for benefits from economic regulation (as we define it) rests on the case of the natural monopoly, where an industry's cost can be minimized only by having a single firm produce. Regulation can then improve social welfare by inducing firms to price efficiently. A case in point is local telephone markets, where the FCC has produced significant welfare benefits to consumers of local phone services through the lower prices that result from mandating interconnections to the local loop. Others contend that natural monopolies are dependent on technology and demand and both have a habit over time of undermining natural monopolies. Furthermore, the history of regulatory agencies which regulate natural monopolies, including the FCC up to the ATT breakup, has been one characterized by their slow adaptation to technological and market changes to the detriment of consumers. Whatever the validity of this generalization, one commenter suggested that currently the benefits of local phone regulation to consumers could be as high as $34 billion. Clearly this issue deserves more study.
Several commenters also raised questions about our estimates of the costs and benefits of environmental regulation. One commenter argued that our base case estimate for the cost of environmental regulation in 1988 was an overstatement because it was based on EPA's Cost of Clean report, which included costs that should have been attributed to state and local regulations. The Cost of Clean report provided an aggregate estimate including the costs of State and local regulations in its totals, but it also disaggregated the total into various components, including an estimate for Federally mandated regulations, excluding costs that arise from State and local actions. We used the Federally mandated costs for our estimates. The commenters also pointed out that some expenditures on automobile emissions regulations would undoubtedly continue even it EPA regulations were rescinded and therefore these costs should not be included. We made this same point in our discussion of the rising baseline phenomenon, but also pointed out that this implied that the benefits of these regulations should not be counted either. There is currently no way to know which expenditures originally incurred because of regulation would continue to be incurred even if those regulations were rescinded. On the other hand, since it is likely that the expenditures that would be continued are likely to be the ones that produce the most benefits, this phenomenon implies that the ex post ratio of benefits to costs is likely to be less than the ex ante ratio. In the report, we generally use ex ante estimates but note that the rising baseline phenomenon implies that both costs and benefits are likely to be overstatements of the effects of removing regulation.
One commenter also asserted that our estimate of the benefits of environmental regulation were understated because the benefit cost ratio that we used to extrapolate benefits for the 1987 to 1996 period, which was based on Hahn's (1996) estimates of the benefits and costs of EPA regulations over the period 1990 to 1995, was based on benefits and costs for those actions and we should have used a benefit-cost ratio based on all regulations, not just the most recent ones. The implicit assumption behind this recommendation is that the benefit cost ratios of environmental regulations systematically decline over time, presumably because the most cost-effective regulations are issued first. If that is so, then using benefit cost ratios for the most recent regulations to extrapolate to all regulations would understate benefits. But there is little empirical evidence for this proposition, perhaps because over time agencies are learning to do a better job in designing more cost-effective regulations. In fact the Hahn and Hird's 1988 baseline benefit cost ratio for environmental regulation is lower that Hahn's (1996) ratio based on regulation issued between 1990 and 1995. Thus the procedure recommended by this commenter would lower the estimated benefits of environmental regulation, not increase them as the commenter asserted.
Direct and Indirect Effects of Regulation
The second issue on which we requested comments was our discussion of the indirect effects of regulation. Several commenters suggested that more work should be done in assessing the indirect impacts of Federal regulation on various sectors of the economy. We acknowledge that our summary of the literature on the direct and indirect effects of regulation on the economy did raise more questions than it answered, but we also believe it is a fair summary of the existing knowledge in the area. Our main points were that these effects are important but they occur on both the cost and benefit side and it is not clear which side predominates. Clearly more work needs to be done on this issue.
One commenter suggested that the affects of regulation on innovation, wages, employment, and income distribution should be examined. We agree. Executive Order 12866 calls on the agencies to examine and consider the distributional and equity effects of regulations that they propose to issue. We believe that both OMB and the agencies can do a better job in estimating the distributional and equity effects of regulation. This is another area that we hope to be able to work on with the agencies over the next few years to improve our performance.
In our July 22nd draft report, we also asked commenters to provide " any additional studies that might provide reliable estimates or assessments of the annual costs and benefits, or direct and indirect effects, of regulation in the aggregate or of the individual regulations" that we discussed in the draft report. Several commenters did suggest, and in some cases sent, studies of the costs and benefits of regulatory programs. Several of the studies have only recently been completed, while others are drafts of reports that are still under review. Moreover, we have not had time to evaluate them for reliability and their consistency with the studies that we have used in arriving at our estimates. Over the next year, we intend to review these materials and any others that become available. Below is a brief description of the studies we received.
The Vice Chairman of the Board of Governors of the Federal Reserve sent us a recently completed staff study that summarizes the results of various other studies of the cost of banking regulation (Elliehausen 1997). The study concludes that the incremental cost of banking regulations in 1991 was about $7.7 billion. Incremental costs was defined as costs that banks would not incur but for Federal regulation. The study also concluded that banks spend another $8 billion or so to comply with regulations that they would spend even in the absence of Federal regulation. One of the studies that the Fed staff relies on is a study by the accounting firm, Grant Thorton (1993), which was sent to us directly by the Independent Bankers Association of America.
The American Water Works Association sent us three studies : Estimating the Costs of Compliance With Drinking Water Standards: A User Guide, Methods for Valuing the Benefits of the Safe Drinking Water Act: Review and Assessment, and the Cost-Effectiveness of SDWA Regulations. This material contains information on cost and benefit estimation methodology and some estimates of the costs of drinking water regulations.
The American Automobile Manufacturers Association sent us several articles, including a 1991 study by NHTSA, Moving America More Safely and Tengs, et al, "Five Hundred Life-Saving Interventions and Their Cost-Effectiveness."
The U. S. Chamber of Commerce sent one of their recent publications, Federal Regulation and its Effects on Business, reporting on two nationwide surveys of business experiences with Federal regulation. The study provides information on how many businesses perceive the regulatory process and offers suggestions as to what types of regulations businesses fine most burdensome.
The Environmental Defense Fund sent their July 1997 report, Toxic Ignorance: The Continuing Absence of Basic Health Testing for Top-Selling Chemicals in the United States, which points out that the benefits of environmental regulation cannot be estimated if we lack basic data on the toxicity of high-volume chemicals.
The American Enterprise Institute sent us two recent studies that it and the Brookings Institution just published, Improving Regulatory Accountability (Hahn and Litan 1997) and An Agenda for Regulatory Reform (Crandall et al 1997). These studies offer numerous proposals intended to improve the regulatory system.
Many other commenters cited various studies in support of their various suggestions to help improve our understanding of the impact of regulation, some of which we were unfamiliar with. We intend to review these studies to broaden our understanding of the issue discussed in the report.
Inefficient Regulatory Programs or Program Elements
Finally, we asked commenters to identify "programs or program elements on which there is objective and verifiable information that would lead to a conclusion that such programs are inefficient or ineffective and should be eliminated or reformed." Many of the recommendations we received were for reforms to the regulatory process. There were surprisingly few specific regulatory programs or program elements that were identified for us. The fact that we received so few proposals for reforming specific programs or regulations, especially accompanied with supportive studies, reinforces our conclusion that it is premature to make specific recommendations about reforming specific regulatory programs. We will continue to review the information we received and other information on the costs, benefits and other effects of regulation for the purpose of finding candidates for reform. The following describes the suggestion we have received to date.
Professor Hopkins of the Rochester Institute of Technology recommended that OMB list particular regulations that "fail" a benefit-cost test, identify the efficiency gain from their reform or rescission, and indicate any legal or other obstacles to taking action. As an example of a rule whose costs exceeds its benefits, Hopkins offered EPA's recently promulgated National Ambient Air Quality Standard (NAAQS) for ozone.
The Business Roundtable (BR) listed both EPA's ozone and particulate matter NAAQSs as "prime examples" of inefficient and ineffective rules, whose costs are ". . . far in excess of any benefits." BR stated that the costs of the standards are so large that the resulting reductions in living standards and subsequent ability to afford health and medical needs would lead to an increase in the incidence of death and serious disease. BR also identified the Corporate Average Fuel Economy standard as a candidate for reform. BR stated this standard ". . . has saved little or no fuel, but, which, when binding, has come at high direct and indirect costs -- including the costs associated with increased highway fatalities and injuries associated with the smaller and lighter vehicles mandated by the standard." BR suggested the use of economic incentives, such as fuel taxes as a more effective and less costly way to reduce fuel consumption.
The U.S. Chamber of Commerce offered the results of two recent surveys of its members. Among the observations were the following: (1) The complexity of pension regulations has effectively prevented a large percentage of firms from offering defined benefit plans; (2) One in six companies laid off workers to deal with the cost of labor and employee benefits regulations and one in three avoided hiring new employees; and (3) One in 10 businesses laid off workers to meet the cost of environmental and natural resource regulations.
The Edison Electric Institute identified EPA's New Source Performance Standards/Prevention of Significant Deterioration programs under the Clean Air Act and the 316 Program under the Clean Water Act as programs that should be reformed to include a benefit-cost approach in determining compliance options.
The National Apartment Association, National Leased Housing Association, and National Multi Housing Association joint comments stated that the analysis supporting EPA's lead-based paint rule was "seriously deficient in many respects." These commenters stated that, given the significant expenditures involved, EPA needs to make clearer the benefits of the program before imposing such costs.
Finally, the American Water Works Association stated that the benefits of the drinking water program are largely attributed to a relatively small handful of contaminants. It recommended EPA maintain sufficient statutory flexibility in selecting contaminants to regulate as well as in the form of those regulations.
In summary, we received many helpful comments from a diverse set of interests. We have much food for thought and much work to do.